Marc, Jennie, Fil, Richard,

Summary: Ruled out problems with QPD B or its in-chamber wiring. Pulled TIA chassis and fixed a loose cable but still not sure what is causing the high noise of QPD B segment 4. Might be time to reconnect the ground pin we pulled on the input cable to see if this somehow caused this problem?

Marc and I tried swapping the pins for segments 4 and 3 at the input to the TIA chassis variant 2 by using a breakout board. This means swapping the anodes (pins 9 and 10) and the cathodes (pins 22 and 23). I re-measured the ASD of QPD B dark noise, where reference 32-35 are today's measurement and 12-15 are the reference measurement from the 24th September. Before and after this change, segment 4 as readout in CDS has a higher noise than the other segments. Since swapping the channels at the TIA input made no change to this we can safely conclude that the problem is not with the in-vacuum equipment and instead with the TIA chassis.

Marc then tried unplugging the connections for both quadrants at the TIA input and, as we have seen previously, the excess noise disappears. References 12-15 are from 24th September and 40-43 are the measurement with the cables unplugged.

Then we pulled the chassis out to see if we could find any obvious problems. Fil and Marc noticed that the cable that connects the PCBs (D1001974) to the output of the chassis was not crimped properly at the inside of the output conenctor. Fil replaced this cable and recrimped the ends.

Marc and I then measured the noise of the chassis in the lab, first with no current sources connected to the input, then with a current source of 1uA at the TIA input.

Every channel had similar noise, with and without the current cource connected.

For no input, every channel has a noise of around 650 nVrms/root Hz at 10 Hz, except for segment 4 which had 900 nVrms/ root Hz. Note this is much smaller than the noise difference between the segments we saw in CDS when the chassis was plugged into the QPDB.

For a 1uA source, every segment has a noise level of around 60-80 uVrms/ root Hz.

We also checked the 4 inputs for the measurement and reference PDs (channels 1-17 at output of chassis) when no current source was attached. and these looked were all a similar level. See the pin in and out diagram here.

After replacing the chassis back in the rack and plugging in all the cables again I remeasured the spectra and did not see much change. The live traces are the measurements from today and refs 12-15 are the dark noise measurement from 24th September.

To sanity check myself I made one pot of various measurements of QPD B segment 4 over time (all with light shuttered and offsets on the QPD filter banks so that the average QPD sum stays the same as whent the shutter is open).

As you can see the noise was worse between 3rd and 4th of September, but only by a small amount. The noise became worse by an order of magnitude after the grounding pin was pulled on the TIA input cable on 16th September (LHO alog #91934).

I am now wondering if us pulling the ground pin on the input cable is what is causing this problem as we did not see it until after we did this. No idea why it would only affect one segment though.